United StatesDepartment ofAgriculture

Forest Service

Pacific NorthwestResearch Station

General TechnicalReportPNW-GTR-510April 2001

Log and Lumber Grades asIndicators of Wood Quality in20- to 100-Year-Old Douglas-Fir Trees from Thinned andUnthinned Stands

R. James Barbour and Dean L. Parry

Authors R. James Barbour is a research forest products technologist and Dean L. Parry is aforester, Forestry Sciences Laboratory, P.O. Box 3890, Portland, OR 97208.

Barbour, R. James; Parry, Dean L. 2001. Log and lumber grades as indicators ofwood quality in 20- to 100-year-old Douglas-fir trees from thinned and unthinnedstands. Gen. Tech. Rep. PNW-GTR-510. Portland, OR: U.S. Department ofAgriculture, Forest Service, Pacific Northwest Research Station. 22 p.

This report examines the differences in wood characteristics found in coastal PacificNorthwest Douglas-fir (Pseudotsuga mensziesii (Mirb.) Franco) trees harvested at theage of 70 to 100 years old to wood characteristics of trees harvested at the age of 40to 60 years. Comparisons of differences in domestic log grades suggest that the pro-portion of log volume in the higher grades (Special Mill and No. 2 Sawmill) increasedwith both stand age and tree size. Simulation of lumber grade yields based on logcharacteristics suggests that yields of higher grades of lumber increased until aboutage 60 to 70, and then leveled off over the rest of the age range examined in thisanalysis. We included structural lumber products in the analysis but not higher valueappearance grade products, and some evidence suggests that yields of these prod-ucts might have begun to increase in the oldest trees. The analysis also showed thatthe younger trees had larger branches and more juvenile wood, possibly becausethey had been grown in stands that received a higher level of early stand manage-ment than the older trees. If these young trees were grown to the ages of 70 to 100,they likely would not produce the same log and lumber grade yields found in the oldertrees we examined.

Keywords: Wood quality, log grade, lumber grade, thinning, Douglas-fir, Pseudotsugamensziesii, ecosystem management, sustainable forestry.

Abstract

Contents 1 Introduction

2 Objectives

3 Methods

4 Source of Data

4 Method of Comparison

5 Field Data Collection

6 Estimated Log Grade Yields

7 Estimates of Lumber Grades

8 Results and Discussion

8 Log Grades

14 Relative Importance of Log Grading Criteria

16 Lumber Yields

18 Conclusions

20 Acknowledgments

20 Literature Cited

1

Introduction In recent years, most landowners in the coastal Pacific Northwest have begun toplace greater emphasis on conserving biodiversity in their management practices.There are, however, important distinctions in the objectives of the various landowners.Whereas most federal lands and some other public lands are now managed primarilyto maintain or restore ecological function, with wood production being a secondaryobjective, wood production is still the primary objective for most private landowners ofindustrial and large nonindustrial land. There are many variations on these broadthemes, but in general, this divergence in objectives results in relatively long rotationsto promote old-forests characteristics on public land and relatively short rotations toenhance the economic returns from wood production on large private land holdings.

Most managers of industrial and large nonindustrial land holdings have adopted standmanagement practices that involve planting and precommercial thinning1 followed byeither a commercial thinning or clear felling by about age 40. The first commercialentry is highly dependent on market conditions. It may replace a precommercial thin-ning at about age 20 if pulp or chip markets are good, or it may not occur until aboutage 40, but in most cases, clear felling usually occurs no later than age 60. Port BlakelyTree Farm in Seattle, Washington, is a moderate size, private owned company thatuses a different strategy. On some of the Port Blakely land, Douglas-fir (Pseudotsugamenziesii (Mirb.) Franco) is managed on 70- to 100-year rotations under silviculturalregimes that may involve multiple thinnings. The regimes used by Port Blakely aresimilar to management activities anticipated for some USDA Forest Service and USDIBureau of Land Management (BLM) land designated as “matrix” or during the periodwhen active management is allowed in “late-successional reserves” under theNorthwest Forest Plan (NWFP) for the recovery of the northern spotted owl (USDAand USDI 1994).2 These management techniques also might interest other landowners,of both large and small parcels, who want to create stand characteristics that differfrom those found in typical industrial plantations but who still consider timber harvestan important land management objective (Curtis and others 1998).

The analysis described here was undertaken as a cooperative project between PortBlakely Tree Farm and the USDA Forest Service, Pacific Northwest (PNW) ResearchStation. We evaluated those differences in wood characteristics likely to exist inDouglas-fir stands managed on shorter (40- to 60-year) and longer (70- to 100-year)rotations. Potential log and lumber quality for trees removed as early as 20 years,which might be expected during early commercial thinning, and trees removed at up

1 Under some market conditions, thinned material is sold for chips,whereas under others it is left on the site.

2 The matrix land allocation under the Northwest Forest Plan forthe recovery of the northern spotted owl (NWFP) “...is the area inwhich most timber harvest and other silvicultural activities will beconducted.” Late-successional reserves are intended to “...maintaina functional, interactive, late-successional and old-growth forestecosystem. They are designed to serve as habitat for late-succes-sional and old-growth related species including the northernspotted owl.”

to 100 years are considered. Economic analysis is not performed in this reportbecause (1) the economics of wood production is not a primary concern under theNWFP, and (2) firms interested in this information will conduct their own economicanalyses using proprietary financial data. This report should, however, provide suffi-cient information about wood characteristics for those interested in conducting eco-nomic analyses.

Data for the analysis came from two sources: (1) new data, collected from a sampleof trees harvested from two Port Blakely thinning trials in naturally regeneratedDouglas-fir stands on site II+ (King 1966) land in southwestern Washington and (2)existing data drawn from a database of tree, log, and lumber characteristics main-tained by the Ecologically Sustainable Production (ESP) of Forest Resources team ofthe PNW Research Station (Stevens and Barbour 2000). The trees used for thisanalysis came from studies conducted by the ESP team in collaboration with a con-sortium of public and private landowners known as the Stand Management Coopera-tive (SMC) (see Fahey and others 1991). These trees are referred to as the SMCtrees.

Curtis (1995) and Stinson (1999) give details about the stand characteristics and thetreatment history for the Port Blakely thinning trials. The older installation, designatedby Port Blakely as XT-1, was established in a 47-year-old stand in 1948 and harvest-ed at age 95. Curtis (1995) characterizes the thinning regime as light. The youngerinstallation, known as XT-7, was established in a 42-year-old stand in 1958 and har-vested at age 80. Curtis (1995) characterizes the treatment of this installation as aheavy thinning. Trees were removed from the Port Blakely thinning trials in summer1996 to determine various wood characteristics. The sample of SMC trees representstrees that are about the same age as the Port Blakely trees. In addition, the SMCtrees include a sample of younger trees intended to provide a comparison withcurrent management practices being used on private land in western Washingtonand Oregon.

The objectives of this analysis were developed to provide information that is useful toboth public and private resource managers interested in understanding the influenceof stand age and management history on wood characteristics. The objectives were to:

1. Make qualitative comparisons of the characteristics of trees grown on PortBlakely’s XT-1 and XT-7 thinning trials to trees in various age classes that repre-sent typical management practice on industrial forest land in western Washingtonand Oregon (SMC trees). We compared:

a. Log and lumber grade yields of thinned and unthinned portions of each XT thinning trial

b. The characteristics of trees grown on the XT thinning trial to those of SMC trees of similar ages

c. The characteristics of both the Port Blakely trees and older SMC trees (80 to 100 years) to those of younger SMC trees (20 to 60 years)

2. Develop a better understanding of the likely wood characteristics of trees grownunder the longer rotation mandated for some lands under the NWPF for matrixland (USDA and USDI 1994).

2

Objectives

3

Methods The three measures of wood quality considered in this analysis are (1) log grade, (2) visual lumber grades, and (3) machine stress-rated (MSR) lumber grades. Thesemeasures were chosen because they can be evaluated directly to compare productpotential of each group, and because price series are readily available for each prod-uct (Anderson 2000, Barr 1998, Warren 1998), thereby making it easy for others toconduct economic analyses by using the product yield information provided in thisanalysis. Only structural products were considered because self-pruning does nottypically occur in Douglas-fir stands until age 60 to 80 (Kachin 1940), and thus eventhe oldest trees used in this analysis had insufficient time to produce appreciableamounts of clear wood.

The available data sets are not ideal for this analysis because many of the trees wereoriginally collected for other purposes. Not all of the records for the trees included inthe analysis contained all the information needed to produce grades under all threesystems. The data sets are, however, adequate to make reasonable estimates of loggrades by using techniques developed by Christensen (1997). Christensen’s (1997)techniques use diameter at each buck point, measurements of branch size by using amodified height pole and caliper, and estimations of growth rate. Simulation of lumbergrade yield is also possible by using regression models developed by Fahey and others(1991). Although estimated log grades and simulated lumber yields do not provide anexact description of wood characteristics in these trees, they do allow reasonablecomparisons among the various age classes and management intensities.

At each of the two Port Blakely installations, 25 sample trees were selected systemat-ically across the full range of diameters present on the unthinned and thinned perma-nent sample plots. This process resulted in selection of 50 trees from each thinningtrial for a total of 100 Port Blakely trees. An additional 289 trees were selected fromtwo earlier SMC studies (Barbour 1995,3 Fahey and others 1991) by using the ESPtree quality database (table 1).

The criteria used for selecting an SMC tree were that (1) tree age (from 21 to 100years) and (2) sufficient data (enough data had to be available about a tree to esti-mate the log or lumber grades used in the analysis). The two previous SMC studiesincluded trees from 20 locations throughout western Washington and Oregon. Theobjectives of these studies and the tree selection criteria were similar in that theyfocused on site class I or II land and trees grown in either natural stands or planta-tions managed for rapid wood volume production. We believe this sample representsthe existing resource on private land, but most Forest Service and BLM land has amuch lower site class. Consequently, site class should always be consideredbefore extrapolating results to federal ownerships. We believe that the similarity insites, management history, and selection criteria made it reasonable to combine treesfrom the two studies. Trees included in these two studies were selected systematicallyto represent the range of variation in diameter and branch size and should not beregarded as random samples. For this reason, generalized inferences should not be

Source of Data

3 Study conducted in cooperation with the SMC.

4

made from analyses presented here. The analyses do, however, provide a qualitativedescription of the way wood characteristics change in different-aged trees and howthe trees from the Port Blakely installations fit into the continuum of changing qualitywith age that can be constructed from the trees in the ESP tree quality database(Stevens and Barbour 2000).

Data were organized into a set of eight analytical groups, the thinned and unthinnedsample from each of the two Port Blakely XT installations and the SMC trees dividedinto four age classes (table 1). The designations used for the Port Blakely trees are95-UT, 95-TH, 80-UT, and 80-TH. They refer to the approximate age at which theinstallation was harvested4 and type of treatment, thinned (TH) or unthinned (UT).The SMC age classes are designated as 21-40, 41-60, 61-80, and 81-100 to reflectthe harvest ages of the trees included in each class.

The treatments at the Port Blakely installations were unreplicated, and the groupingof the SMC trees is by age class rather than by stand. Although artificial, these group-ings are thought to represent the population of Douglas-fir managed under similarconditions. The response variables for this analysis were generated either by combin-ing empirical data on the morphological characteristics of logs or by using regressionmodels developed elsewhere to estimate lumber grades. Consequently, comparisonsdescribed in this analysis are based on interpretations of tables and graphs of theaverages for each group rather than statistical analysis of the variation among thegroups.

4 Based on estimated stand initiation date. Average stump ringcounts for the sample trees are shown in table 1.

Method of Comparison

Table 1—Source, naming, average age, and sample size of trees used to makeup the analytical groups compared in this analysis

Source of trees and Treatment or Mean analytical groupa age class (years) stump age Number of trees

YearsPort Blakely:

95-UT XT-1 (UT) 91 2595-TH XT-1 (TH) 93 2580-UT XT-7 (UT) 72 2580-TH XT-7 (TH) 73 25

Stand Management Cooperative:

21-40 21-40 32 12441-60 41-60 49 11361-80 61-80 64 4181-100 81-100 86 11

a UT = untinned; TH = thinned.

5

Field Data Collection

The Port Blakely trees were compared by installation and treatment. The type ofgrouping used for the SMC trees was chosen because commercial thinnings tend tofall into the 21- to 40-year age class, second thinnings and clear felling on privateland tends to coincide with the 41- to 60-year age class, and one of the Port Blakelyinstallations falls into each of the next 20-year age classes. On national forests andBLM land, the final thinning in late-successional reserves will occur between ages 60and 80, and harvests from the “matrix” land could fall into any of these classes(USDA and USDI 1994).

When the Port Blakely trees were felled and bucked, each log was labeled with aunique number that identified the treatment (TH or UT), tree from which the log came,and position the log came from within the tree. Stump ring counts were taken todetermine tree age. Woods length logs were moved to a paved log yard where addi-tional measurements were made including large-end diameter, small-end diameter(SED), and midpoint diameter inside bark, as well as diameter inside bark 4 feet fromthe large end of each butt log. Large-end, small-end, and juvenile wood diameter foreach log were measured according to standard scaling procedure, which uses theaverage of the smallest cross-sectional diameter and a diameter measured at 90° tothe smallest diameter. Midpoint diameter of the Port Blakely logs was measured out-side the bark. Bark thickness was measured to the nearest 0.1 inch and subtractedfrom the diameter outside the bark to obtain diameter inside the bark. Other measure-ments included bark thickness at each buck point, number of rings on the large andsmall ends of each log, branch index (BIX) for each mill length log, and juvenile corediameter on the large and small ends of each woods length log. Branch index wascalculated as the average of the largest diameter knot in each of the four quadrants(90°s of circumference) of the log surface. The juvenile wood core was estimated asthe diameter of the first 20 growth rings from the pith on each end of each woodslength log. This method was used to estimate juvenile wood because it is a requiredinput for the Fahey and others (1991) MSR lumber grade yield models.

Port Blakely trees were bucked to preferred woods length logs of 35 feet to an 8-inchtop. Stand Management Cooperative trees were bucked into 40-foot woods lengthlogs with a minimum top diameter of 4 inches. Lengths differed because Port Blakelymanufactured their logs for export, whereas the SMC trees were processed for thedomestic market. Analysis of log grades is based on woods length logs. Analyses oflumber grade yields are based on mill length logs. The Port Blakely logs were soldfull length so data for the lumber analysis was taken on “pencil-bucked” mill lengthlogs. This was generally one 20- and one 14-foot log except when a short woodslength log was produced because of a falling break, some other defect, or short toplogs. Stand Management Cooperative woods length logs were bucked into mill lengthlogs and sawn into lumber (Fahey and others 1991; Barbour and others [see footnote3]). In both SMC studies, the preferred mill length log was 20 feet, but shorter logsalso were produced when woods length logs were not full length or contained abreak or defect.

6

Log grades were not recorded in the field, but estimated grades were developed foreach woods length log. Methods for log grade assignment basically follow thosedeveloped by Christensen (1997) for estimating log grade in standing trees.Christensen’s (1997) methods are similar to those pioneered by Gaines (1962) andrefined by Lane and Woodfin (1977), except that Christensen (1997) includes actualmeasurement of branch size in the lower portion of the stem and growth rate atbreast height. Differences in the methods used in this analysis included directmeasurement of branch size and growth rate for all logs because the trees werefelled, which made log ends and branch stubs accessible. Estimated grades werebased on the small-end diameter, branch size, growth rate in the outer portion ofthe log, and minimum length (table 2).

Growth rate in the outer portion of the log was estimated for both Port Blakely andSMC woods length logs by subtracting the radius of the juvenile core (first 20 ringsfrom the pith) from the total radius and dividing the result by the number of rings inthe mature wood portion of the log (equation 1). This was done for both ends of thelog and averaged. Although this is not exactly how growth rate is defined in the loggrading rules (Northwest Log Rules Advisory Group 1998), it is sufficient to obtainan adequate estimate of the growth rate (GR ),

GR = Total ring count - 20, (1)rm

where

Total ring count is the number of rings in the large or small end of the mill length log,20 is the number of rings in the juvenile core, andrm is the radius of the mature wood portion of the log calculated as the radius of thelog minus the radius of the juvenile core.

Estimated Log Grades

Table 2—Criteria used to estimate log gradesa

Log grade

No. 2 No. 3 No. 4Grading criteria Special Mill Sawmill Sawmill Sawmill

Minimum rings per inchouter portion 6 rings per inch None None None

Minimum scaling diameter 16 inches 12 inches 6 inches 5 inchesMaximum branch diameter 1.5 inches 2.5 inches 3 inches NoneMinimum length 17 feet 12 feet 12 feet 12 feeta These criteria follow the Northwest Log Rules Advisory Group (1998) rules except for the branch frequencyrequirements for Special Mill and net scale requirements for all grades. Failure to include the net scalerequirement is most important for No. 4 Sawmill logs because in practice, larger logs with high defect levelsare placed in this grade. In this analysis, only small logs or those with large branches were graded as No.4 Sawmill.

7

The maximum branch diameter was taken as the largest of the four measurementsmade on each woods length log segment for calculating BIX. The Special Mill gradehas an additional branch frequency requirement (Northwest Log Rules AdvisoryGroup 1998, p.11). This feature could not be included in this analysis because nobranch frequency data were collected. We may therefore have overestimated thenumber of Special Mill logs.

An SAS program (SAS 1995) was written to sort and grade logs. There was insuffi-cient information to grade either peelers or No.1 Sawmill logs. The protocol was tofirst sort logs by diameter (those 16 inches and greater for Special Mill) then by knotsize. Those that met both the diameter and knot size restrictions for Special Mill werethen tested to ensure that growth rate was less than six rings per inch and that logsmet the appropriate length requirement. Logs failing to meet requirements for SpecialMill were tested as No. 2 Sawmill logs and so on (table 2).

No adjustment in log grade was made for net scale because the logs from the PortBlakely trees were not scaled. Field observations revealed little defect or deformationin these logs, but because the defects were not recorded, they could not be includedin the log grading protocol. The methods used to assign grade restricted the No. 4Sawmill grade to logs with an SED of less than 6 inches. In reality, larger defectivelogs, with low net scale volumes, also are placed in this grade. This illustrates a short-coming in the analysis because some defective larger logs could in theory have beengraded as other than No. 4 Sawmill. On the other hand, even the oldest stands in thisstudy are relatively young in terms of the biological capability of Douglas-fir. Stands inthis age range typically would not display high levels of defect. The PNW ResearchStation over the past 30 years has analyzed many logs from several Douglas-fir woodproduct recovery studies to evaluate the importance of the net scale requirement.This evaluation was accomplished by using the database described by Stevens andBarbour (2000). The results suggest that the average defect in trees younger than100 years old is always less than 5 percent of the gross scale volume and usuallymuch less than that. So we feel that the grading system used here adequately repre-sents log grade for the purposes of this analysis.

Estimates of visual and MSR lumber grade yields for each mill length log (8 to 20feet) from both the Port Blakely and the SMC trees were made by using grade recov-ery models developed by Fahey and others (1991). These estimates are based on theBIX for visual grades and BIX and juvenile wood percentage for MSR grades.

Selection of lumber grades—The visual lumber grades reported in this analysisare those included in the structural joists and planks for 6-inch and wider lumber orStructural Light Framing grades for 4-inch wide lumber (WWPA 1991). The MSR lum-ber grades are 2100 fb, 1650 fb, 1450 fb, Utility (or No. 3 depending on the width ofthe lumber), and Economy (WWPA 1991). This set of grades is not the same one thata mill would choose when processing Douglas-fir. Mill managers typically select a mixof grades that maximizes their value recovery. The optimization process includesinformation about current markets, orders, and the technical capability of the mill. Thegrade recovery models in Fahey and others (1991) attempt to generalize a set ofgrades useful for describing the characteristics of a resource.

Estimates of LumberGrade Yields

8

In practice, mills that visually grade lumber under the Structural Light Framing or theJoists and Planks rules also use other visual rules to increase value. Sometimes vari-ous laminating grades are sorted to capture the increased value of higher qualitylumber. In other cases, grading rules are mixed to minimize production of low-valuegrades. For example, 2 by 4 lumber that would be downgraded to No. 3 and sold foras little as $180 per thousand board feet5 under the Structural Light Framing rulesoften qualifies as Standard under the Light Framing Rules (WWPA 1991) and can besold for $335 per mbf. This ability to mix grading systems could make a considerabledifference in the profitability of a mill.

The situation for MSR lumber is somewhat different. Fahey and others (1991) choseto represent MSR lumber as a set of three MSR grades and two visual grades. Inreality, mills produce a much more complex mix of grades. Grading rules allow visualrestricted grading options for lumber that fails to make the lowest MSR grade manu-factured by a mill. The rules require placing this lumber in visual grades with mechani-cal properties below the lowest MSR grade the mill produces. The 1450 fb gradeestimated by the Fahey and others (1991) model has design values rated lower thanany Douglas-fir visual grade except Utility. If a Douglas-fir mill actually produced the1450 fb grade, it would preclude manufacture of all other visual grades except Utilityand Economy. For the same reasons given in the example for No. 3 mentioned earlier,it is unlikely Douglas-fir mills actually would produce the mix of MSR grades reportedin this analysis. This grade mix does, however, provide a good profile of the distribu-tion of properties within the resource.

The data used to discuss quality at the log level are presented in three tables. Table 3contains information on the proportion of log volume in each grade for each analyticalgroup. Table 4 includes data about the relative importance of morphological charac-teristics (diameter, knot size, and growth rate) and processing decisions (length) indetermining log grade. These data also are used to demonstrate how some log orstumpage purchasers may benefit by going beyond log grade when making process-ing decisions. Table 5 summarizes details on the relative importance of each grade-determining characteristic.

Port Blakely thinning trials—All four Port Blakely analytical groups produced goodquality logs. Special Mill and No. 2 Sawmill were the two predominant log grades onall four areas. The Port Blakely installations yielded almost no No. 4 Sawmill logsbecause the utilization standards set by the company (8-inch minimum SED) effec-tively precluded their manufacture under the grading protocol used in this analysis.A small volume of No. 4 Sawmill logs was manufactured from the 95-UT trees, butotherwise the lowest grade was No. 3 Sawmill (table 3).

Results andDiscussion

5 Based on prices reported by Anderson (2000) from September2000.

Log Grades

9

Table 3—The proportion of total log volume in each analytical group meetingvarious log grades displayed by analytical group

Log grade Stand averageb

Analytical Special No. 2 No. 3 No 4 quadratic mean groupa Mill Sawmill Sawmill Sawmill Age diameter

- - - - - - - - - - Proportion - - - - - - - - - - Years Inches

95-UT .41 .45 .13 .01 91 24.980-UT .41 .44 .15 .00 72 22.695-TH .56 .36 .08 .00 93 26.980-TH .40 .54 .06 .00 73 29.321-40 .00 .20 .74 .06 32 14.541-60 .04 .38 .53 .05 49 19.761-80 .03 .48 .48 .00 64 16.781-100 .37 .38 .19 .06 86 22.4

a UT = unthinned; TH = thinned.b Quadratic mean diameter at breast height of sample trees.

Table 4—Reason why logs were not placed in the next higher log grade byanalytical group and log grade

AverageTotal grading

Grading criteria grading criteriaAnalytical Log Growth Total criteria exceededgroupa grade Diameter Knots rate Lengthb logs exceeded per log

- - - - - - - - - - Numbersc - - - - - - - - - -

95-UT No. 2 Sawmill 25 17 0 0 36 42 1.1780-UT No. 2 Sawmill 29 17 4 0 39 50 1.2895-TH No. 2 Sawmill 29 22 1 2 38 54 1.4280-TH No. 2 Sawmill 22 36 31 1 50 90 1.8021-40 No. 2 Sawmill 14 8 13 1 14 36 2.5741-60 No. 2 Sawmill 39 45 24 0 55 108 1.9661-80 No. 2 Sawmill 20 7 1 1 23 29 1.2681-100 No. 2 Sawmill 9 3 0 0 10 12 1.20

95-UT No. 3 Sawmill 24 0 0 0 24 24 1.0080-UT No. 3 Sawmill 31 0 0 0 31 31 1.0095-TH No. 3 Sawmill 17 0 0 0 17 17 1.0080-TH No. 3 Sawmill 14 0 0 0 14 14 1.0021-40 No. 3 Sawmill 137 8 0 0 138 145 1.0541-60 No. 3 Sawmill 151 23 0 0 156 174 1.1261-80 No. 3 Sawmill 58 1 0 0 58 59 1.0281-100 No. 3 Sawmill 10 1 0 0 10 11 1.10

95-UT No. 4 Sawmill 0 1 0 0 1 1 1.0080-UT No. 4 Sawmill 0 0 0 0 0 0 095-TH No. 4 Sawmill 0 0 0 0 0 0 080-TH No. 4 Sawmill 0 0 0 0 0 0 021-40 No. 4 Sawmill 22 2 0 0 23 24 1.0441-60 No. 4 Sawmill 12 4 0 0 16 16 1.0061-80 No. 4 Sawmill 0 0 0 0 0 0 081-100 No. 4 Sawmill 1 1 0 0 1 2 2.00

a UT = unthinned; TH = thinned.b Length depends on merchandising decisions by individual companys.c Entries represent number of times this grading criteria was exceeded.

10

Both thinned areas had a somewhat better log grade distribution (more Special Milland No. 2 Sawmill volume) than the corresponding unthinned areas. About 85 percentof the volume from both unthinned areas was in Special Mill and No. 2 Sawmill gradelogs, whereas the thinned areas yielded 94 percent (95-TH) and 92 percent (80-TH)in these grades. Although no precise records are available, the strategy was to thinfrom below and attempt to remove mortality and defective trees (Curtis 1995). Thispractice presumably improved log grade in the residual stand, thereby eliminating thetrees that would later yield low-grade logs. Stinson (1999) reports data for the volumeremoved from the XT-1 (95-TH here) installation during the various entries but not thequality of this material.

The log grade distributions in the two unthinned areas were virtually identical, but thelog grade distributions in the thinned areas differed from one another, thereby sug-gesting that the intensity of the thinning and possibly harvest age may have affectedlog grade distribution. The Port Blakely 95-TH trees were older at harvest and receiveda lighter thinning treatment than the trees on the 80-TH area. Both of these factorsare consistent with the observed log grade yield results. Proportional volume of SpecialMill grade logs was about one-third higher for the 95-TH analytical group than any ofthe three other Port Blakely analytical groups (fig. 1). Although it is not possible to usethe available information to say conclusively why these differences occurred, sufficientinformation exists to rule out some possibilities and highlight others. The most likelyexplanations for these types of differences are (1) tree size, (2) spacing at stand initi-ation, (3) timing of thinning, (4) thinning intensity, and (5) age at harvest. Of these,thinning intensity seemed the most likely cause for the differences in log gradeobserved for the Port Blakely trees.

Table 5—Proportion of logs downgraded from Special Mill to No. 2 Sawmill orNo. 3 Sawmill by analytical groupa

Grading criteriaAnalytical Log Growthgroup grade Diameter Knots rate Length

95-UT No. 2 Sawmill 0.69 0.47 0.00 0.0080-UT No. 2 Sawmill .74 .44 .10 .0095-TH No. 2 Sawmill .76 .58 .03 .0580-TH No. 2 Sawmill .44 .72 .62 .0221-40 No. 2 Sawmill 1.00 .57 .93 .0741-60 No. 2 Sawmill .71 .82 .44 .0061-80 No. 2 Sawmill .87 .30 .04 .0481-100 No. 2 Sawmill .90 .30 .00 .00

95-UT No. 3 Sawmill 1.00 .00 .00 .0080-UT No. 3 Sawmill 1.00 .00 .00 .0095-TH No. 3 Sawmill 1.00 .00 .00 .0080-TH No. 3 Sawmill 1.00 .00 .00 .0021-40 No. 3 Sawmill .99 .06 .00 .0041-60 No. 3 Sawmill .97 .15 .00 .0061-80 No. 3 Sawmill 1.00 .02 .00 .0081-100 No. 3 Sawmill 1.00 .10 .00 .00

a Proportions sum to greater than 1.0 because some logs exceed more than one grade determining criteria.

11

Tree size is not the reason. The 80-TH trees were larger yet yielded a lower proportionof volume in Special Mill (table 3), the most diameter-sensitive grade (table 2). Thestands also were thinned at about the same age, so it is unlikely that age at thinningled to the observed difference in grade yield.

Spacing at stand initiation is important because it is one of the major factors deter-mining branch size (Larson 1969). Little information is available about the conditionsin these stands at the time they were established, but Curtis (1995) concluded thatboth Port Blakely installations were probably initially widely spaced. Curtis bases thisconclusion on the number of trees present when the stands were first thinned aswell as evidence of mortality before thinning. Although it is unknown if the stocking inone stand was heavier than in the other, the closeness of proportional grade yieldsfor the unthinned areas supports the idea that the stocking levels were similar andthat the thinning treatment, not initial conditions, created the differences betweeninstallations. Curtis also rates both sites as II+ so differences in site quality wereprobably not a factor.

Thinning intensity may have influenced log grade yield. The 95-TH area was thinnedfrequently but lightly at 47, 55, 60, 74, and 77 years old. The 80-TH area was thinnedheavily at ages 42 and 54 then lightly at age 62. The relative density of the 80-THarea had not recovered to a point where suppression-related mortality was anticipatedby the time it was remeasured at age 76 (Curtis 1995). Data presented in table 4show that growth rate and knot size were much more important in limiting log gradefor the logs from the 80-TH analytical group than for those from any of the other PortBlakely analytical groups. This is consistent with expectations for stands maintained inrelatively open conditions.

SM

2S

3S

4SPro

port

ion

of t

otal

Average tree age and treatment

80-UT 80-TH 95-UT 95-TH

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0

Figure 1—Proportion of total woods length log volume in each log grade from each Port Blakely analyticalgroup (UT = unthinned; TH = thinned). SM = Special Mill, 2S = No. 2 Sawmill, 3S = No. 3 Sawmill, and4S = No. 4 Sawmill.

The differences in treatment appear to have promoted rapid branch and stem diame-ter growth in the trees from the 80-TH area but only moderate branch growth andstem diameter growth in the trees from the 95-TH area. Information collected on BIXsupports this idea even though it was not possible to reconstruct histories of branchsize over time. Change in branch size with tree height provides a surrogate for thisinformation. Lower branches provide a picture of conditions before treatment, where-as the upper branches provide a snapshot of treatment effects. The average BIX foreach 17.5-foot log from each Port Blakely analytical group is plotted in figure 2.Although it is not possible to say which logs grew before or after the thinning, it is rea-sonable to generalize the influence of thinning from the height trends. On both instal-lations, BIX was smaller in the lower logs from the thinned areas than the lower logsfrom the unthinned areas, thereby suggesting that either these differences existedbefore thinning or some trees with large branches were removed during thinning. Thelatter is consistent with the goal of improving both growth and quality of the residualtrees by removing small, poorly formed, and branchy trees. Regardless of the causeof this initial difference, the height trend indicates a response to thinning. As logheight increased, the BIX on both thinned areas became larger than the BIX on thecorresponding unthinned areas. The lighter thinning treatment received by the trees inthe 95-TH area seems to have increased stem diameter enough to improve yield ofSpecial Mill logs without increasing either branch size or growth rate enough to dis-qualify many of those larger logs from the higher grade. In contrast, the heavier thin-ning received by the trees in the 80-TH area increased both branch diameter and ringwidth beyond the allowable limits, thereby disqualifying many large logs from theSpecial Mill grade (table 5).

Trees from the 95-TH area were 15 years older at harvest than those from the 80-THarea, but it is unclear if age was an important factor in the observed differences in loggrade yield. Proportional log grade yields from the two unthinned areas were virtuallyidentical (table 3). Trees from the younger 80-TH area were larger (table 3), yet yieldedabout one-third less of their total volume in Special Mill grade logs than the older95-TH trees (fig. 1). This happens because growth rate and branch size were moreimportant factors in determining grade in the younger trees than in the older ones(table 4). Extending the rotation without additional thinning could result in slowerdiameter growth, which in turn might allow more logs to meet the growth rate criteriafor Special Mill. The 95-TH trees were also old enough to begin branch stub occlusionafter natural pruning (Kachin 1940). Occlusion also might result in reduction in num-ber of knots counted by the scaler during grading because they become less obviousas they are overgrown. Using the methods developed by Christensen (1997) for pro-jecting stand developmental trajectories might help to evaluate the importance ofthese potential changes and answer questions about how the characteristics of the80-TH trees might change were they grown for another 15 years.

Stand Management Cooperative trees—Age does, however, seem to explain muchof the change in proportional log grade when the full 20- to 100-year age range isexamined. The proportion of No. 3 Sawmill log volume declined with age, whereas theproportion of volume in No. 2 Sawmill and Special Mill logs increased over this agerange (fig. 3). An exception to this trend was seen in the age class 41 to 60 years and61 to 80 years where there was a slight decline in the proportional volume of Special

12

13

Figure 2— Change in branch index (BIX) with mill length log height (17.5 feet per log) for the trees fromeach Port Blakely analytical group.

Bra

nch

inde

x (in

ches

)B

ranc

h in

dex

(inch

es)

Log number (no. 1 = butt)

Log number (no. 1 = butt)

A

B

Mill logs (table 3). This is apparently linked to the diameter of the two groups (propor-tion of logs downgraded to No. 2 sawmill for diameter, table 5). The trees from the 41-to 60-year age class had a quadratic mean breast height diameter (QMD) of 19.7inches. The trees from the 61- to 80-year age class had a smaller QMD of only 16.7inches. The Special Mill grade requires a minimum SED of 16 inches, so in at leasthalf of the trees in the 61- to 80-year age class, a 34-foot butt log was unlikely toyield a Special Mill log. This decline in the proportion of Special Mill logs occurredeven though the younger 41- to 60-year age class trees had an average BIX that wasnearly one and one-half times as large as the 61- to 80-year age class trees (1.41inches vs. 0.97 inch). Thus, the No. 2 Sawmill logs produced from the younger 41- to60-year age class trees would be noticeably “rougher” than those from the 61- to 80-year age class (table 4). This is evident from the higher number of grading criteriaexceeded per log. Although it is impossible to completely separate diameter from agein this analysis, diameter was always an important factor in determining log grade inall analytical groups. All but 33 of the 362 logs from the SMC database graded as No.3 Sawmill, and all those from the Port Blakely thinning trials placed in that grade weredowngraded only for diameter (table 4).

Table 4 contains information about the reasons logs failed to qualify for the next high-est log grade. Even though two groups of logs may be assigned to the same grade,their “quality” can be quite different. The latitude in quality allowed within a single gradeis apparent when the number of grading criteria exceeded per log is compared amongall analytical groups for the No. 2 Sawmill grade (table 4, final column in rows 1-8).

14

Pro

port

ion

of t

otal

vol

ume

Average group age, age class, and group average diameter

21-4

0/32

41-6

0/49

61-8

0/64

81-1

00/8

6

0.8

0.7

0.6

0.5

0.4

0.3

0.2

0.1

0

SM

2S

3S

4S

Figure 3—Proportion of total woods length log volume in each log grade for analytical groups of SMC treesarranged by age class (x-axis label also shows average age and quadratic mean diameter for the tree ineach group). SM = Special Mill, 2S = No. 2 Sawmill, 3S = No. 3 Sawmill, and 4S = No. 4 Sawmill.

Relative Importance of LogGrading Criteria

The logs from trees in the younger two SMC age classes exceeded more gradingcriteria than either those from the older SMC age classes or the Port Blakely trees.The 95-UT No. 2 Sawmill logs (row one, last column in table 4) exceeded an averageof 1.17 grading criteria per log, and those from SMC age class 21-40 exceeded anaverage of 2.57 grading criteria per log (row 5, table 4). This means that for the 95-UTtrees, about 17 percent of the No. 2 Sawmill logs contained features that exceeded atleast two grading criteria. The other 83 percent of the No. 2 Sawmill logs only exceed-ed one grading criterion. The No. 2 Sawmill logs from the SMC age class 21-40 typi-cally exceeded two grading criteria, but 57 percent of them exceeded three. All ofthese logs were nonetheless assigned to the No. 2 Sawmill grade under existingrules (Northwest Log Rules Advisory Group 1998). This means that at least theappearance of the logs from the two groups would be noticeably different. Some dif-ferences in the types of products manufactured from them also might be expected,and in fact, the market currently recognizes the range of quality within some loggrades because different prices are reported for old- and young-growth Special Mill,No. 2, and No. 3 Sawmill grades (Log Lines 1999).

More subtle but possibly important differences also are seen among the Port Blakelyanalytical groups. For example, the data presented in table 3 suggest that, based onlog grade alone, the trees from thinned areas have higher quality than those fromunthinned areas. When the data are reexamined by using the techniques presented intable 4, a different, more complex picture develops. Although the trees from thinnedstands produce larger proportional volumes in the higher grades, the No. 2 Sawmilllogs from the thinned stands exceed more grading criteria per log than those from theunthinned stands. This suggests that at least by some measures, the logs within theNo. 2 Sawmill grade from the unthinned stands are of higher quality. This result isconsistent with more general information on branch size and quality seen after thin-ning reported for both Douglas-fir (Maguire and others 1991, Riou-Nivert 1989) andother species (Alazard 1994, Barbour and others 1994, Colin and others 1992).Consequently, more demanding log grading systems, such as those used for propri-etary export grades, might reveal greater differences between trees from the thinnedand unthinned area, than are apparent here.

Comparison of Port Blakely trees and SMC database trees—The proportion oflogs that exceeded each grading criterion for each analytical group is shown in table5. This representation of the data shows the influence of the heavy thinning receivedby the 80-TH trees on the criteria used to evaluate log grade. The 80-TH analyticalgroup is the only Port Blakely analytical group where growth rate was an importantfactor in disqualifying logs from the Special Mill grade. Knot size was also moreimportant for this group than for the other three Port Blakely analytical groups. Basedon the data in table 5, the No. 2 Sawmill logs from the 80-TH trees are most similar tothose from the SMC 41- to 60 year age class. Or, in other words, No. 2 Sawmill logsfrom this 80-year-old stand might closely resemble those from stands one half to two-thirds that age. It is important to keep in mind, however, that although the logs gradedas No. 2 Sawmill from these two groups had similar characteristics, 40 percent of thevolume from the 80-TH group qualified as Special Mill, but only 4 percent of the vol-ume from the age class 41- to 60 year SMC trees made that grade (table 3).

15

Estimated yields of visually graded and MSR lumber for the trees from the PortBlakely thinning trials and the SMC trees are presented in tables 6 and 7. These arecomposite results for all logs from all trees in each analytical group. Lumber gradeyields are reported as a percentage of the total board-foot lumber tally for each ana-lytical group. Lumber tally is dependent on log diameter and was estimated by usingthe lumber recovery factor model of Fahey and others (1991).

16

Table 6—Average volume percentage of yields of visually graded lumber bygrade from each analytical groupa

Lumber gradeAnalytical Select Utility or No. 2 and Branchgroup Structural No. 1 No. 2 No. 3 Economy Better. index

- - - - - - - - - - - - - - - - Percent - - - - - - - - - - - - - - - - Inches

95-UT 21 20 46 10 4 86 0.8980-UT 17 21 46 11 5 84 .9095-TH 25 17 45 10 4 87 .8780-TH 17 19 48 12 5 83 1.0521-40 5 15 53 19 8 73 1.3241-60 5 12 51 22 11 67 1.4161-80 16 20 47 12 5 83 .9781-100 25 21 43 9 4 88 .88a Lumber grade yields were estimated by using grade prediction models developed by Fahey and others(1991) for WWPA Structural Light Framing and Joists and Planks grade rules (WWPA 1991).

Table 7—Average volume percentage of yields of machine stress rated (MSR)lumber grades for each analytical groupa

Lumber gradeAnalytical Utility or Juvenilegroup 2100fb 1650fb 1450fb No. 3 Economy wood

- - - - - - - - - - - - - - - - - - - - Percent - - - - - - - - - - - - - - - - - - - -

95-UT 25.0 37.2 18.0 13.4 6.3 28.280-UT 23.3 36.9 18.2 14.9 6.6 35.095-TH 24.9 36.5 20.0 12.3 6.3 24.180-TH 22.5 35.7 18.3 16.0 7.4 24.821-40 9.1 26.0 24.1 29.4 10.7 83.341-60 8.6 23.5 21.6 29.1 17.2 57.561-80 22.0 37.3 16.3 15.9 8.5 39.081-100 28.8 40.7 13.2 10.5 6.8 33.8a Grades were estimated by using equations developed by Fahey and others (1991) for MSR lumber.The "grades" are approximations based on edge knot size and stiffness.

Lumber Yields

The average BIX for each analytical group is included in table 6 because the visuallumber grade yield models developed by Fahey and others (1991) use BIX as theironly independent variable. These relations are independent of log diameter. Branchindex varied in the narrow range between 0.89 and 1.05 inches for all the analyticalgroups except the SMC age classes 21 to 40 years and 41 to 60 years, which hadmuch larger branches (BIX 1.32 and 1.41 inches). As mentioned earlier, these treesprobably grew in stands that were either established at wider spacings or respacedearlier than either the older SMC stands or the Port Blakely thinning trials.

The recovery models for MSR lumber include both BIX and percentage of juvenilewood as independent variables (Fahey and others 1991) and so percentage of juve-nile wood is included in table 7. The MSR grade yield models have a slight depen-dence on diameter because the juvenile wood diameter is fixed once trees reach 20years old at a given height. As total diameter increases, the proportion of juvenilewood slowly declines, and MSR grade yield improves.

Port Blakely trees—The results for visual lumber grade yield were consistent withthe log grade results already presented. Estimates of visual lumber grade yield for thetrees from all the Port Blakely thinning trials were similar regardless of whether thearea was thinned, although the older (95-TH and 95-UT) trees had slightly higheryields of Select Structural lumber. The results for MSR lumber agreed with those forvisually graded lumber (table 7). The similarity between trees from the thinned andunthinned areas indicates that when middle-aged stands are thinned, there is littleinfluence on quality if structural lumber is the intended product. This result is consis-tent with Christensen’s (1997) findings for the Hoskins level-of-growing-stock installa-tion. The Port Blakely thinning trials were thinned when stands were 40 years old orolder and, by this time, the branches in the lower portions of the stems, where thegreatest volume is concentrated, were almost certainly dead. Branch size was there-fore fixed so faster growth after thinning would not necessarily substantially alterlumber grade distributions.

Stand Management Cooperative trees—The results for visually graded and MSRlumber for the SMC database trees are different than those for the Port Blakely thin-ning trials (tables 6 and 7). The distribution of both visual and MSR grades indicates aconsiderable change in quality in the trees that are older than 60 years as comparedto those younger than 60 years. At this level of analysis, it is not clear if these resultsreflect some quirk in the data or if these differences actually represent differences inquality related to changing silvicultural practices over the past 40 years. Initial spacingor precommercial thinning seems to have influenced the characteristics of the SMCdatabase trees where the younger trees have considerably larger branches (table 6)and higher percentage of juvenile wood (table 7). The observed increases in branchsize and proportion of juvenile wood are consistent with results reported in the litera-ture for widely spaced and early thinned Douglas-fir and other species (Barbour andothers 1997, Briggs and Smith 1986). It therefore seems likely that the information intables 6 and 7 actually reflects differences in young stand management. This is notsurprising because the SMC trees were intentionally selected to represent treesgrown by using silvicultural practices designed to favor rapid wood production.

17

Conclusions

If this assumption is correct, these data indicate that the practices used to grow theyounger trees have put them on a trajectory for lower quality that will influence theircharacteristics and consequently their market value under any economically viablerotation. Even if the trees from the SMC age classes 21 to 40 and 41 to 60 yearswere grown until they reached age 80, they still would have large branches in theirlower boles as well as higher in the stem if rapid growth is maintained during theentire rotation. The BIX reported for these trees in table 6 is fixed and will not declinewith age. The relation between growth rate and the proportion of juvenile wood iscomplex because at any height, the volume of juvenile wood is fixed once trees makethe transition to mature wood production. The longer the trees grow, therefore, theless juvenile wood they will contain as a proportion of their total volume. Given twotrees of equal size, the one that grew more slowly when young will tend to have alower proportion of juvenile wood than the one that grew more quickly when young.Although the proportion of juvenile wood will slowly decline with increased diameter,large branches will still negatively influence MSR lumber grade yields. The importantpoint is that simply growing the younger trees longer would not lead to the sameproduct yields as are seen in the trees from the Port Blakely thinning trials or the twoolder age classes of SMC trees.

Management history seemed to influence the characteristics of the trees examinedhere. Analysis of the SMC trees by age class reveals a gradual increase in log andlumber grades with increasing stand age. The Port Blakely trees fit into this continuumfairly well. Log grade, visual lumber grade, and MSR lumber grade results suggestthat Douglas-fir wood quality increases fairly rapidly until about stand age 60 and thenapparently levels off. The trees from the five oldest analytical groups had similar logand lumber grade yields although the SMC trees in this age class came from thecentral Cascade Range in Oregon, and the Port Blakely thinning trials were on theOlympic Peninsula in western Washington. The evaluation of lumber grades performedin this analysis did not, however, recognize increasing clear wood volume becauseonly structural lumber was considered in the estimated grade yields. Data available inthe literature on natural pruning suggest that quality may begin to increase again atabout age 100 when recoverable amounts of clear wood have formed. The log graderesults tend to account for this because the Special Mill grade is intended to capturethe potential yield of Select and Better lumber. Even so, there was little difference inlog grade between the older and younger Port Blakely thinning trial that could beattributed to age.

The differences in quality between trees from thinned and unthinned Port Blakely areaswere relatively small. The trees from thinned areas tended to yield larger proportionsof higher graded logs than from unthinned areas. For logs of equal grade, however,logs from the unthinned areas tended to have fewer characteristics that preventedthem from being placed in the next higher grade than the logs from the thinned areas.So, some processors might consider logs of equal grade from unthinned areas ashigher quality than the logs from thinned areas. There were essentially no differencesin estimated grade yields of visually graded or MSR structural lumber among thetrees from the various analytical groups from the Port Blakely thinning trials. Thisanalysis did not, however, include estimates of grade yields for appearance gradelumber, which might have revealed greater differences in quality between the logsfrom thinned and unthinned areas.

18

19

Diameter was always an important factor in determining if logs were placed in allgrades; however, two of the other grading criteria, that is, knot size and growth rate,were as important or more important than diameter in determining if logs met thecriteria for the Special Mill grade. Even when young trees are large enough to makeSpecial Mill logs, they rarely do because either their branches are too large or theirgrowth rate is too high. Branch size and growth rate also are related to stand tending;for example, initial spacing and the timing and intensity of thinning, as well as sitequality.

A continuum of improving log and lumber grade yield was apparent with tree age, andthis was common across analytical groups. For example, the trees from the SMC ageclass 81 to 100 years yielded similar proportions of logs and lumber in the variousgrades as did the Port Blakely trees. As the age of the SMC trees declined, branchsize and juvenile wood dimensions changed in ways that could not be accounted forsimply by tree or stand age, and these changes also influenced log and lumber grade.This probably reflects a shift from more extensive to more intensive forestry practiceson industrial land in western Oregon and Washington over the past 50 years, but inmost cases, we have insufficient plot level treatment histories to say this conclusively.Practices such as use of genetically improved planting stock, herbicides to reducevegetative competition in young stands, wide initial spacing, precommercial thinning,and commercial thinning all are consistent with the trend toward the larger branchesand wider juvenile wood cores observed in the younger trees. It is, for example, clearfrom the analyses presented here that trees grown by using the long rotation forestryand relatively late thinnings practiced in Port Blakely’s XT-1 and XT-7 thinning trialshad different wood characteristics than trees grown by using the high-yield forestrytechniques practiced for the two youngest SMC age classes and that those differ-ences would persist even if the SMC trees were grown to the same age as the PortBlakely trees.

The potential importance of site quality, however, should be considered when inter-preting the results from this study. All of the trees analyzed here came from eithersite I or site II land, and yet much of the land owned by the federal government is athigher elevations and lower site classes where slower growth is expected even withearlier and heavier thinnings. This might tend to mitigate branch size, growth rate,and juvenile wood diameter, but it would reduce total diameter, thereby making theeffect on log grade yield difficult to predict.

These results do, however, suggest that the longer rotations used under the NWFP“matrix” land specifications could result in trees with different wood characteristicsthan those grown either in industrial plantations or of older trees grown under pastforest management practices. This, of course, depends on the timing and intensity ofthinnings used to alter structure in the existing stands. The earlier and heavier thethinnings, the more the trees from longer rotations will resemble trees from plantationsmanaged for rapid growth. If late heavy thinnings like the regime used on the PortBlakely XT-7 installation (80-TH analytical group) are used, large trees with moderatebranch size and growth rates should result. The positive effect of diameter will lead tosome improvement in the domestic log grade mix from the older stands but little, ifany, change in the proportions of various lumber products manufactured from them.On the other hand, late light thinnings, similar to those used by Port Blakely on theXT-1 thinning trials (95-TH analytical group), will result in trees that are somewhat

Acknowledgments

smaller for a given age but with characteristics that might make them more desirablefor certain wood products with higher market value. If early, heavy thinnings are used,the negative effect on proportional log and lumber grades probably will be more pro-nounced than that seen for the two Port Blakely installations. These differences stillneed to be quantified, but it is apparent from the difference in knot size, juvenile woodproportion, and growth rate between the more intensively managed SMC trees andthe Port Blakely trees that longer rotations that result in larger trees will not necessar-ily result in trees with significantly higher market value for wood removed from thesestands.

We thank all of the technical experts who reviewed and commented on this report andwho provided information used in conducting the study. We are particularly grateful toMike Mosman and Garret Eddy from Port Blakely Tree Farm who provided partialfunding for the study and many technical comments; Steve Stinson from the Universityof Washington, who assisted with fieldwork and provided data from his own study;and Sue Willits, Bob Curtis, Dean DeBell, Dave Marshall, and Glenn Christensen allfrom the PNW Research Station, and Jim Stevens from the Campbell Group whoprovided technical input and reviewed the manuscript at various stages.

Alazard, P. 1994. Stand density and spacing of Pinus pinaster: consequences forgrowth and tree quality. Moulis-en-Medoc, France Afocel Sud-Ouest. Informations-Foret,-Afocel-Armef. 2: 129-144.

Anderson, J.P. 2000. Random Lengths: the weekly report on North American forestproducts markets September 8, 2000. Eugene, OR: Random Lengths: 55(1): 5.

Barbour, J. 1995. The Stand Management Cooperative annual report. In: Woodquality project progress report. Seattle, WA: University of Washington, College ofForest Resources: 20-25.

Barbour, R.J.; Fayle, D.C.F.; Chauret, G. [and others] 1994. Breast-height relativedensity and radial growth in mature jack pine (Pinus banksiana) for 38 years afterthinning. Canadian Journal of Forest Research. 24(12): 2439-2447.

Barbour, R.J.; Johnston, S.; Hayes, J.P.; Tucker, G.F. 1997. Simulated stand char-acteristics and wood product yields from Douglas-fir plantations managed forecosystem objectives. Forest Ecology and Management. 91: 205-219.

Barr, L.K. 1998. Pacific Rim wood market report no. 132. Gig Harbor, WA: Wood NotePublishing.

Briggs, D.; Smith W.R. 1986. Effects of silvicultural practices on wood properties. In:Oliver, C.D.; Hanley, D.P.; Johnson, J.A., eds. Douglas-fir: stand management forthe future. Contribution No. 55. Seattle, WA: University of Washington, College ofForest Resources, Institute of Forest Resources: 1-8-117.

Christensen, Glenn A. 1997. Development of Douglas-fir log quality, product yield,and stand value after repeated thinnings in western Oregon. Corvallis, OR: OregonState University. 176 p. M.S. thesis.

Colin, F.; Monchaux, P.; Houllier, F.; Leban, J.M. 1992. Growth, branching andwood quality of Norway spruce: first results for trees harvested at a third thinning.Paris, France. Association Foret-Cellulose (AFOCEL); Annales-de-Recherches-Sylvicoles,-AFOCEL,-1991: 251-296.

20

Literature Cited

Curtis, R.O. 1995. Extended rotations and culmination age of coast Douglas-fir: oldstudies speak to current issues. Res. Pap. PNW-RP-485. Portland, OR: U.S. Depart-ment of Agriculture, Forest Service, Pacific Northwest Research Station. 49 p.

Curtis, R.O.; DeBell, D.S.; Harrington, C.A. [and others]. 1998. Silviculture for mul-tiple objectives in the Douglas-fir region. Gen. Tech. Rep. PNW-GTR-435. Portland,OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest ResearchStation. 123 p.

Fahey, T.D.; Cahill, J.M.; Snellgrove, T.A.; Heath, L.S. 1991. Lumber and veneerrecovery from intensively managed young-growth Douglas-fir. Res. Pap. PNW-RP-437. Portland, OR: U.S. Department of Agriculture, Forest Service, PacificNorthwest Research Station. 25 p.

Gaines, E.M. 1962. Improved system for grading ponderosa pine and sugar pine sawlogs in trees. Tech. Pap. 75. Berkeley, CA: U.S. Department of Agriculture, ForestService, Pacific Southwest Forest and Range Experiment Station.

Kachin, T. 1940. Natural pruning in second-growth Douglas-fir. Res. Note 31. Portland,OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Forest andRange Experiment Station: 5.

King, J.E. 1966. Site index curves for Douglas-fir in the Pacific Northwest. ForestryPaper No. 8. Centralia, WA: Weyerhaeuser Corporation.

Lane, P.H.; Woodfin, R.O., Jr. 1977. Guidelines for log grading coast Douglas-fir.Res. Pap. PNW-218. Portland, OR: U.S. Department of Agriculture, Forest Service,Pacific Northwest Research Station. 14 p.

Larson, P.R. 1969. Wood formation and the concept of wood quality. Bulletin No. 74.New Haven, CT: Yale School of Forestry. 54 p.

Log Lines. 1999. Log lines log price reporting service September 1999 prices.Mt. Vernon, WA: 2(10): 1-6.

Maguire, D.A.; Kershaw, J.A., Jr.; Hann, D.W. 1991. Predicting the effects of silvicul-tural regime on branch size and crown wood core in Douglas-fir. Forest Science.37(5): 1409-1428.

Northwest Log Rules Advisory Group. 1998. Official log scaling and grading rules.8th ed. [Place of publication unknown]. 48 p.

Riou-Nivert, P. 1989. Wood quality, pruning and silviculture of Douglas fir. RevueForestiere Francaise. 41(5): 387-410.

SAS. 1995. SAS procedures guide, version 6, 3d ed., 5th printing. Cary, NC: SASInstitute. 705 p.

Stevens, J.A.; Barbour, R.J. 2000. Managing the stands of the future based on thelessons of the past: estimating western timber species lumber recovery using his-torical data. Res. Note. PNW-RN-528. Portland, OR: U.S. Department ofAgriculture, Forest Service, Pacific Northwest Research Station. 9 p.

21

Stinson, S.D. 1999. 50 years of low thinning in second growth Dougals-fir. TheForestry Chronicle. 75(3): 401-405.

U.S. Department of Agriculture and U.S. Department of the Interior. 1994. Recordof decision for amendments to Forest Service and Bureau of Land Managementplanning documents within the range of the northern spotted owl. [Place of publica-tion unkown]. 74 p. [plus Attachment A: standards and guidelines].

Warren, D.D. 1998. Production, prices, employment, and trade in Northwest forestindustries, second quarter 1997. Resour. Bull. PNW-RB-228. Portland, OR: U.S.Department of Agriculture, Forest Service, Pacific Northwest Research Station.130 p.

[Western Wood Products Association]. 1991. Western Lumber Grading Rules 91.Portland, OR. 238 p.

22

The Forest Service of the U.S. Department of Agriculture isdedicated to the principle of multiple use management of theNation’s forest resources for sustained yields of wood, water,forage, wildlife, and recreation. Through forestry research,cooperation with the States and private forest owners, andmanagement of the National Forests and National Grasslands,it strives–as directed by Congress–to provide increasinglygreater Service to a growing Nation.

The United States Department of Agriculture (USDA) prohibitsdiscrimination in all its programs and activities on the basis ofrace, color, national origin, gender, religion, age, disability,political beliefs, sexual orientation, or marital or family status.(Not all prohibited bases apply to all programs.) Persons withdisabilities who require alternative means for communicationof program information (Braille, large print, audiotape, etc.)should contact USDA’s TARGET Center at (202) 720-2600(voice and TDD).

To file a complaint of discrimination, write USDA, Director,Office of Civil Rights, Room 326-W, Whitten Building, 14thand Independence Avenue, SW, Washington, DC 20250-9410or call (202) 720-5964 (voice and TDD). USDA is an equalopportunity provider and employer.

Pacific Northwest Research Station 333 S.W. First Avenue P.O. Box 3890 Portland, OR 97208-3890

Pacific Northwest Research Station

Web site http://www.fs.fed.us/pnwTelephone (503) 808-2592Publication requests (503) 808-2138FAX (503) 808-2130E-mail desmith@fs.fed.usMailing address Publications Distribution

Pacific Northwest Research StationP.O. Box 3890Portland, OR 97208-3890

U.S. Department of AgriculturePacific Northwest Research Station333 S.W. First AvenueP.O. Box 3890Portland, OR 97208-3890

Official BusinessPenalty for Private Use, $300